Industry Practices Is Your Company a Candidate for Predictive Dialing Technology? This article is based on a presentation given by Cara Taussig of PERT Survey Research and Kathleen FitsGerald of CRI at the CASRO Marketing Research Technology Conference in New York, New York. Predictive dialing differs from autodialing in general in that it “dials ahead” of a group of interviewers who may, or may not, be available to take the next live contact reached. Other autodialing modes such as power- preview- or proactive dial at a ratio of 1:1 (dialing only if an interviewer is available). The objective of the predictive dialing algorithm is to match the next interviewer available to take a call with the next “live” respondent it reaches as quickly as possible within the parameters that you set. While attempting to reach a live respondent, predictive dialers disposition busies, no answers, fax/computer tones and Standard Information Tones (SIT) without interviewer intervention. Marketing Research Adapting Predictive Cautiously In late 1994, only 50% of field services and 75% of full-service research services employed CATI. Only 3-4% of research firms employed predictive dialing which was used by 74% of telemarketers. More affordable PC-based and LAN-based predictive dialers were first introduced in the mid-90's. Predictive dialing is an important tool available to researchers to address the changing call-center environment. How will predictive dialing be combined with other processes and tools to enable telephone research to remain feasible throughout this decade? For example, will it be used for just-in-time panel building for internet research, for sample building through inbound/outbound customer care centers, or for pre-recruiting for central location interviewing? The Problem or the Solution for Quality Research As we substitute technology for labor in the call center, we must carefully weigh many competing factors and predictive dialing is one piece of the puzzle. Effective market research is based on a combination of cost, quality and timeliness. We are used to technology helping us maintain or improve one or more of these factors. But how do we react when our technology helps one of these at the possible expense of another? How do we achieve the right balance? Predictive dialing is such a technology. It promises production increases and, hence, better profits. We are all under tremendous pressure to produce results more quickly and cheaply, and are faced with an increasingly challenging environment. But our industry is justifiably concerned with respondent cooperation. Predictive dialing may be one factor impacting respondents’ willingness to cooperate. Respondent cooperation is integral to having a non-biased, projectible sample which is at the heart of research quality. Further, respondent cooperation is at the heart of the long-term survival of our industry. With the proliferation of phone numbers being used for machines such as answering machines, faxes, modems, and pagers as well as the proliferation of phone numbers and multiple phone numbers per household, predictive technology enables less elapsed time to find “good” numbers compared with manual or power dialing. There is high interviewer turnover because of other opportunities and because interviewing is a difficult job with uncooperative respondents. High turnover leads to unprofessional interviewers. Interestingly, this is an area where predictive dialing is also part of the solution. Because predictive dialers handle so many tasks for interviewers, they allow more rapid “on the job” training speaking with respondents. Similarly, researchers were faced with low unemployment rates and rising labor costs. Again, this is an area where predictive is part of the solution because it maximizes the completes per hour that an interviewer obtains during a shift. However, predictive dialing is not a panacea and does not resolve all inhibitors facing market research. Predictive can even raise issues such as call abandonments where the respondent answers the phone only to find no one on the other end and pauses which occur after the respondent says hello and before the interviewer speaks. Although you can control both abandonment rates and interviewer response times with predictive, there remains the possibility of an annoyance factor. There is also legislation being written that could affect the use of dialers. Faced with these inhibitors, we may feel overwhelmed trying to balance our commitment to maintaining respondent cooperation with our pressing business need to be profitable and timely in the production of quality market research. Lessons Learned from Other Researchers We spoke at length with more than 10 market research and CATI firms, all using well-known predictive dialer equipment. Here are some of the common themes they expressed: Expect installation delays. Telecomm issues were the most frequently mentioned installation problem. The tip from other companies - get your phone lines installed and tested early. Predictive dialing is a big project requiring coordination with your local phone provider, your long distance provider, your CATI supplier and your predictive dialer hardware and software provider. Second installations are always smoother. There are PBX integration issues. Integrating with your PBX may be possible, you’ll need to weigh the added complexity and additional costs against your needs. Some dialers offer silent monitoring, but when unavailable, some companies have replaced the PBX monitoring functionality with Melco units. Depending on the predictive dialer equipment you choose, you may need to manage the routing of inbound phone calls to manned non-dialer booths. One company double-cabled their predictive dialer booths and installed two-line phones - one line for the dialer, the other for the PBX. That allows the interviewer to choose which telephone hardware they need depending on whether it’s an inbound or outbound phone call. Expect 20-25% production increases or more. Nearly everyone enjoys at least a 20-25% production increase. The highest mention was 60% and the lowest was 10%. Everyone is very concerned about abandonment. Responses to the question “What abandonment rate are you using”? ranged from 0% to 3% with most being 1% or less. The most important dialer selection factor is integration with CATI. Integration with their CATI system was rated extremely important by every company interviewed. One of the key features of this integration is the ability to use quota control and sample disposition tallies available through the CATI software. Use references. Use references to verify delayed response, answering machine detection, CATI software integration, and the strengths and weaknesses of support. Start small and grow. Prove your ROI and develop technical and sample management skills on a small scale and then increase your booths. A powerful technology Predictive dialing is a powerful tool to help market researchers provide timely, high quality, and cost effective market research information. However, the market research industry must continue to lead the way in using this technology responsibly by keeping abandoned calls as low as possible and cleansing do-not-call numbers from their sample. Back to Sawtooth News…